On 22nd August 2018, ESA launched the first direct detection Doppler wind lidar into space. Operating at 355 nm and acquiring signals with a dual channel receiver it allows wind observations in clear air and particle polluted regions of the atmosphere. Measuring with 87 km horizontal and 0.25-2 km vertical resolution between ground and up to 30 km in the Stratosphere the global coverage of Aeolus observations shall fill gaps in the global observing system and thus improve numerical weather prediction.

After first promising results soon after launch, calibration and validation of Aeolus products becomes an essential part of the mission. In order to determine and understand observation biases and random errors as well as possible instrument degradation independent measurements of co-located instruments are necessary.

As part of the German contribution to the Aeolus validation, ground-based wind and aerosol measurements will be used from tropospheric radar wind profilers, Doppler wind lidars, radiosondes and aerosol lidars. Especially radar wind profilers using frequencies between 400 MHz and 500 MHz with their ability to measure wind with high accuracy in clear air as well as in regions with particles and clouds up to 16 km are very well suited for Aeolus comparisons. With the German weather service operating 482 MHz radar wind profilers at four sites in Germany on a 24/7 basis, enough data for long-term statistics will be gathered when Aeolus is orbiting the Earth on a weekly repeat cycle on an sun-synchronous orbit.

In a comparison of instruments with different spatio-temporal resolution and measurement geometry sensing different volumes of the atmosphere, special attention must be paid to the representativeness of both measurements. Methods and techniques must be developed to quantify these sampling errors in order to make a statement about the uncertainties of Aeolus observations by using co-located measurements.

In this presentation, the method for comparing radar wind profiler measurements with Aeolus observations in terms of representativeness is described and demonstrated on selected cases.

In the frame of the German initiative EVAA (Experimental Validation and Assimilation of Aeolus observations), Cal/Val activities for Aeolus are performed by the Ludwig Maximilian Universität Munich (LMU), the Leibniz Institute for Tropospheric Research (TROPOS) and the German Weather Service (DWD).
As part of these activities, direct validation measurements take place at TROPOS in Leipzig, Germany (51.35 N, 12.43 E). The measurements comprise a multiwavelength-Raman-polarization lidar and a Doppler wind lidar. Thus, Aeolus level 1b, 2b, and 2a will be evaluated. While the aerosol information (particle backscatter coefficient, particle extinction coefficient, and particle depolarization ratio) are available within the whole troposphere and lower stratosphere, the Doppler wind lidar measurements are restricted to the local planetary boundary layer (PBL) and lofted aerosol layers. Thus, focus of the evaluation will be put on the aerosol products and PBL winds of Aeolus. Observations at Leipzig have been performed since the very beginning of the Aeolus mission. An example of such an early campaign measurement is shown in Fig. 1 and 2, which present the aerosol and wind observations during the overpass (with respect to the ground track of Aeolus) on 15 of September 2018. As during the time of writing AEOLUS CAL/Val has not yet started, only the ground-based measurements are shown.

In addition to the observations in Germany, long-term measurements at Punta Arenas (53.13 S, 70.88 W), Chile, will start end of November 2018 in the frame of the DACAPO-PESO campaign (dacapo.tropos.de). The Leipzig Aerosol and Cloud Remote Observations System (LACROS) is employed there at the southern tip of South America. LACROS comprises a similar suite of instruments and thus complements the northern hemispheric validation of Aeolus level 1b, 2a, and 2b data at Leipzig.
In this contribution, we will show first results from our observations at Leipzig and the Southern hemispheric location of Punta Arenas.

Fig 1. Observations with the aerosol lidar at Leipzig, Germany on 15 September 2018. Left: Temporal development of the range-corrected signal at 532 nm. Red rectangle is the Aeolus overpass time frame (+- 1.5 hours). Orange dashed frame is the analyzed period for the retrieval of the backscatter coefficient. Right: Particle backscatter coefficient obtained with the Raman method.

Fig. 2. Wind observations (direction indicated by arrows) obtained from scanning Doppler lidar at Leipzig during Aeolus overpass on 15 September 2018.

The Met Office (UK) will be evaluating Horizontal Line of Sight (HLOS) winds from ESA's Aeolus satellite as part of the Calibration and Validation phase of the satellite mission. As part of this assessment, we will compare HLOS winds from Aeolus to our numerical weather prediction (NWP) models, and to other wind observations. We will monitor the HLOS winds routinely from the start of the Cal/Val period, in order to build up long term statistics. We will also look at the HLOS winds using case studies, such as tropical cyclones. Using our NWP model enables us to generate statistics globally, and throughout the whole HLOS wind vertical range.

Here we will present our assessment of the quality of the Aeolus HLOS winds, and their suitability for assimilation into our NWP models. Using observation minus background (O-B) statistics, we will investigate the HLOS winds for biases, and characterise any measurement errors. This investigation will help to inform the quality control strategy for assimilation of HLOS winds. We will present some first results of assimilation of HLOS wind into our global NWP model, using 4-Dimensional variational assimilation (4D-VAR), and our assessment of the impact of their assimilation on the analysis and weather forecasts. Error correlation statistics may also be investigated, using O-B and observation minus analysis (O-A) techniques. Future plans for assimilation of Aeolus HLOS winds and improvements to our HLOS wind assimilation scheme will be discussed.

ECMWF has since the 1990’ies been closely involved in defining Doppler Wind Lidar mission requirements for weather prediction and improving the understanding of the climate. This has right from the early phase been about the instrument design and the scientific and meteorological/climate aspects of the Aeolus Doppler Wind Lidar mission. ECMWF has since 2004 also been involved in the development of the L2B processor and is now part of ESA’s Ground Segment. This keen interest at ECMWF has been, and still is, due to the high importance of direct wind observations for numerical weather prediction (NWP) and to improve the understanding of, primarily, the tropical general circulation. Despite the significant increase in satellite data and in situ observations over the last two decades, there is still a very strong requirement for direct wind profile observations. So, the NWP community expects Aeolus will help us fill this gap in the global observing system. This presentation will present the preliminary investigation on how the Aeolus Doppler Wind Lidar influences the NWP forecasts and the general circulation in the ECMWF NWP system. Special focus will be on comparing the tropical circulation and how it interacts with the extra-tropical weather systems when Aeolus wind data is/isn’t assimilated. As it is still early days for assimilation of Aeolus data at ECMWF, the presentation will focus on more qualitatively aspects of potential improvements in the representation of the general circulation in the tropical/extra-tropical region.
The Aeolus L2B processor provides observation error estimates based purely on the Aeolus measurements. It is a clear advantage if these estimates can be made accurately purely based on the data itself. We will provide an analysis of how these observations error estimates compare with Desroziers statistics and innovation statistics from data assimilation experiments that use Aeolus winds in the ECMWF NWP system.

A prototype new system has been developed in CMA to derive AMVs based on dual-GEO stereo view. In this new system, wind information is derived by using multi-geostationary satellite stereoscopic observations. the motion information is derived by optical flow technique, and the wind height information is derived by parallax technique. With these methods, not only the motion of cloud but also dust/smoke can be obtained. This can to some extent solve the problems of the cloudless dust/smoke target height assignment in the current operational AMV which is only based on IR/WV two bands information. At present the system has been applied on the CMA geostationary series satellites located at 79E,99E,105E and 112E. the early evaluation of the AMVs from this system based on the CMA geostationary satellite data suggests the RMSE is 566meters and Bias is 384 meters on average. During current AMVs validation in 2018, CMA research group find AMVs from this system have some systematic bias over Tibet plateau high mountain and Gobi-desert region. The system is planned to be upgraded in early 2019 to support IODC satellite observation mission , such as EuMetsat MSG at 41.5E and CMA FY-2H at 79E, which will provide better observations over Tibet plateau high mountain and Gobi-desert region.The ESA Aeolus mission provide the unique information for bias diagnose and develop the algorithm to eliminate the AMVs bias. In this presentation, the intercomparsion of Aeolus wind and RAOB will be discussed at first , then the stereo view AMV generation system and its validation with Aeolus wind record will be reported. All the study above will be focused in Northwest China. if there are dust and/or smoke happens in spring 2019 as well as both Aeolus and Geosynchronous satellites have asynchronous observations , we will make some experiments to eliminate bias caused by the time asynchron of different observations source, the AMV accuracy sandstorm scenario in NorthWest China will be discussed in details.

In recent years, aerosol prediction has come to the forefront thanks also to advances in aerosol observability. Several centres with forecasting capabilities now offer forecasts of aerosol fields up to five days ahead. At ECMWF, the Copernicus Atmospheric Monitoring Service (CAMS) delivers aerosol products based on analysis of Aerosol Optical Depth (AOD) from various sensors. Recent research efforts have also shown the potential of lidar backscatter to extract aerosol information. In particular, ESA has funded a collaboration between ECMWF and MétéoFrance to look at the potential of lidar data from the ALADIN instrument on board of Aeolus. While this instrument is designed for wind retrieval and not specifically to do aerosol science, some useful information can be extracted. This is also preparatory work for the EarthCARE mission which will operate as well a UV lidar at the wavelength of 355 nm, designed to study aerosols and clouds. In the context of the A3S project, simulated Aeolus lidar backscatter has been successfully assimilated in the ECMWF 4D-Var. Synthetic datasets have been created from a run of the Modèle de Chimie Atmosphérique de Grande Echelle (MOCAGE) model which provided aerosol fields for the study period (10-20 June 2012). Cloud fields have been incorporated from a run of the Integrated Forecasting System (IFS). These datasets have been utilised to run the Aeolus End-to-End simulator (E2S) to generate synthetic lidar returns from which
the L2A particle backscatter product was retrieved using the official algorithm developed at MétéoFrance. An additional synthetic dataset based on Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO) data was also created. The CALIPSO lidar operates at a different wavelength (532nm) and the orbital parameters are also different from those of Aeolus. The CALIPSO data were processed at KNMI and rescaled at the operating frequency of the Aeolus lidar. The orbital parameters were also adjusted and the E2S simulator provided Aeolus-like data that were fed into the L2A retrieval suite. Both datasets were processed at ECMWF and assimilated into the aerosol analysis using a
lidar backscatter operator developed for this purpose. In view of the recent successful launch of the Aeolus satellite in August 2018, ECMWF will have early access to the data which will be also tested in the assimilation system. This work presents results from the A3S project as well as preliminary work on the use of actual Aeolus particle backscatter data.